¹Université Libre de Bruxelles, Institut de Recherche Interdisciplinaire en Biologie Humaine et Moléculaire, and Laboratory of Medical Chemistry, Erasme Hospital, Brussels, Belgium

Correspondence: Dr. Jean-Marie Boeynaems, Chimie Médicale, Hôpital Erasme, 808 Route de Lennik, 1070 Bruxelles, Belgium. E-mail: jmboeyna@ulb.ac.be

Abstract

In this issue, Seiffert et al. show that adenosine triphosphate enhances the production of inflammatory mediators by human dermal endothelial cells. However, a growing literature shows that adenosine triphosphate exerts anti-inflammatory effects, partly by inducing a semi-maturation of dendritic cells associated with immune tolerance. These disrepancies can be reconciled knowing that extracellular nucleotides constitute danger signals that are released in response to both external aggression (chemical, microbial) and excess inflammation.

Extracellular nucleotides constitute danger signals ... released in response to both external aggression ... and excess inflammation.

In this issue, Seiffert et al. (2006) show that adenosine 5'-O-(3-thiotriphosphate) (ATPS) and other nucleotides exert a proinflammatory effect on cells of the immortalized dermal cell line HMEC-1 (human microvascular endothelial cell-1). ATPS enhanced the release of IL-6, IL-8, and monocyte chemoattractant protein-1 and increased the surface expression of ICAM-1. These effects might favor the recruitment of leukocytes in the skin. Furthermore, the authors show that the spontaneous secretion of these mediators is decreased in the presence of suramin, a nonspecific inhibitor of various nucleotide receptors, thus suggesting an autocrine action of nucleotides. In view of their pharmacological profile, these actions are probably mediated by several receptors; one of them is likely the P2Y₂ receptor, as the action of adenosine triphosphate (ATP) and ATPS was partially reproduced by uridine triphosphate (UTP).

Seiffert et al. (2006) add a new and significant piece to the growing evidence that extracellular nucleotides can play a proinflammatory role, involving the expression of adhesion molecules and secretion of various cytokines and chemokines. Indeed, Seye et al. (2004) have recently demonstrated in another model of vascular endothelial cells, human coronary endothelial cells, that uridine triphosphate acting on the P2Y₂ receptor upregulates the expression of vascular cell adhesion molecule-1 (VCAM-1) via transactivation of vascular endothelial growth factor receptor-2. Uridine diphosphate, acting on the P2Y₆ receptor, has been shown to stimulate the production of IL-8 by both human monocytic cells (THP-1) (Warny et al., 2001) and human lung epithelial cells (A549) (Khine et al., 2005). Furthermore, it was shown in these two models that the release of uridine diphosphate and autocrine activation of P2Y₆ amplify the IL-8 secretion induced respectively by lipopolysaccharide in monocytic cells (Warny et al., 2001) and by human neutrophil peptides in lung epithelial cells (Khine et al., 2005). In human macrophages, ATP promotes the transcription of the IL-6 gene as a result of cytosolic calcium increase (Hanley et al., 2004). Similar stimulatory effects of ATP on IL-6 have also been observed in human osteoblasts and thyrocytes. Lipopolysaccharide-stimulated monocytes produce large amounts of intracellular and biologically inactive IL-1 that is processed and released after activation of the P2X₇ receptor by ATP; this mechanism is deficient in subjects with a loss-of-function polymorphism of P2X₇ (Sluyter et al., 2004). So various nucleotides (such as ATP and UTP) acting on different receptors (such as P2Y₂, P2Y₆, and P2X₇) in different cell types (such as endothelial cells, monocytes/macrophages, and epithelial cells) stimulate the expression of adhesion molecules (such as ICAM-1 and VCAM-1) and the secretion of a range of cytokines and chemokines (such as IL-1, IL-6, IL-8, and monocyte chemoattractant protein-1).

From a physiological standpoint, nucleotides can be viewed as danger signals that are released from cells inter alia in response to chemical or microbial aggression. For instance, ATP is released from keratinocytes in response to various chemical irritants, and the resulting skin inflammation is exacerbated in mice deficient in CD39, an ecto-nucleoside triphosphate diphosphohydrolase that plays a key role in the degradation of extracellular ATP and adenosine diphosphate (Mizumoto et al., 2002). Nucleotides are also released from epithelial cells after invasion by bacterial (McNamara et al., 2001) or viral (Kunzelmann et al., 2004) pathogens. However, it must be emphasized that mechanisms other than cell damage and necrosis can lead to the appearance of nucleotides in the extracellular fluid, such as vesicular trafficking and exocytosis from platelets and neurons, so that the role of extracellular nucleotides is definitely not restricted to that of a danger signal.

It is now realized that the proinflammatory action of extracellular nucleotides is only one part of the story. Like Janus, they have two faces and exert anti-inflammatory as well as proinflammatory actions. The best demonstration of this duality comes from dendritic cells. Dendritic-cell maturation, in response to agonists of Toll-like receptors such as lipopolysaccharide or CD40 ligand, involves the loss of endocytosis, the surface expression of the major histocompatibility complex and of costimulatory molecules, and the secretion of inflammatory cytokines such as IL-12 that plays a key role in the differentiation of CD4⁺ T cells into T helper type 1 cells. ATP mimics some of these actions, in particular the surface expression of costimulatory molecules (such as CD80 and CD86) (Wilkin et al., 2001). However, at the same time it inhibits the secretion of inflammatory cytokines, such as IL-12 (Wilkin et al., 2002), and inflammatory chemokines, such as monocyte chemoattractant protein-1 (Horckmans et al., 2006), this last effect being exactly opposite to what Seiffert et al. (2006) observed in endothelial cells. This semi-maturation of dendritic cells by ATP appears to be mediated by the P2Y₁₁ receptor, which is dually coupled to cAMP and calcium increases (Wilkin et al., 2001). Other mediators that increase cAMP, such as prostaglandin E₂, induce a similar semi-mature state, but the action of ATP displays some unique features. In particular, it induces a much larger secretion of thrombospondin-1 than does prostaglandin E₂ (Marteau et al., 2005). Thrombospondin-1 can inhibit the proliferation of CD4⁺ T cells via interaction with CD47. These and other data suggest that the action of ATP on dendritic cells might favor a state of immune tolerance. In monocytes also, ATP inhibits the secretion of inflammatory cytokines (tumor necrosis factor-) and chemokines (monocyte chemoattractant protein-1) (Kaufmann et al., 2005). A direct immunosuppressive effect of ATP and other nucleotides on CD4⁺ T cells has also been described; it involves the inhibition of both T helper type 1 (IL-2, IFN-) and type 2 (IL-5, IL-10) cytokines (Duhant et al., 2002).

The shift from pro- to anti-inflammatory action of nucleotides might be related to the concentration reached. The observation that low concentrations of ATP induce a recruitment of immature dendritic cells, whereas high concentrations of ATP inhibit their migratory capacity through P2Y₁₁ activation, constitutes a nice example of concentration-dependent effects of nucleotides on inflammation (Schnurr et al., 2003). One might also speculate that this duality is related to the site of nucleotides' release. Their action on endothelial and epithelial cells would stimulate leukocyte recruitment, whereas their direct effect on recruited dendritic cells and lymphocytes could inhibit further recruitment.

Interestingly, whereas inflammation induced by irritant chemicals was exacerbated in CD39^-/- mice, allergic contact hypersensitivity was severely attenuated (Mizumoto et al., 2002). Although the exact mechanism of this defect remains unclear, this discrepancy between two models of skin inflammation fits well with the duality of nucleotide action. Taken together, these data suggest a revised model of danger-signal action of nucleotides. On one hand, their release in response to chemical or microbial aggression signals the need to develop an inflammatory response. However, nucleotide release from damaged and necrotic cells also results from an excessive inflammatory response that needs to be moderated.